A decade of commits

Plug-and-play car telematics — an OBD dongle sold to individuals, micro-fleets, large fleets and insurers. I built backend services and worked across the Android and iOS apps: theft and tow alerts first, then diagnostics and a score for how you had been driving. It went out through partnerships with Uber and ICICI Lombard, past 10,000 devices installed.

Firmware across successive device generations: our own boards in the early years, later CAN-bus reads and factory-fit OE integrations.

The IoT stack behind Bounce’s last-mile scooter fleet in Bangalore — around 25,000 vehicles. You book a scooter in the app and the platform unlocks it remotely: security hardware and immobiliser on the vehicle, rider authentication, billing, and the services that keep every asset findable. At its peak it was taking upward of 10 million server pings a day.

The engine that turns raw device data into something useful: trip reconstruction, GPS trajectory correction, and area computation from geospatial data. Two of the patents came out of this.

The telematics cloud everything else runs on, built deliberately hardware-agnostic: it speaks to most IoT hardware, so bringing a new device platform on board is a quick integration rather than a rebuild. Around it sits the whole customer lifecycle — device manufacturing, SIM management, installation, aftersales support and subscription renewals.

Retrofit telematics for tractors and harvesters — live tracking, fuel monitoring, acreage, trip replay and asset security, with the apps owners and operators actually open. Shipped as the Simha Kit, later rebranded as the Krish-e Smart Kit.

Much more than the Smart Kit. The Krish-e app itself — crop calendar, agronomy advisory, and photo lo, rog jaano, where a photograph of an ailing plant comes back with a diagnosis — in nine languages and past a million installs. Alongside it: Krishi Diary, the farm bahi-khata for logging work, expenses and mandi prices; the CRM behind the call centre; demo-plot tracking; and the data lake under all of it.

Decision analytics for sugar mills. Satellite imagery and field data predict sugar content and yield field by field, so a mill knows when each block is at its sweetest and should be cut — timing that is the difference between an average season and a profitable one. Seven seasons in, with 30+ mills across India, more than 200,000 farms analysed and a cumulative 350,000 acres under satellite analysis. I own this one end to end — the architecture, the prediction models and their evaluation, the web tooling, and the cloud operations underneath.

A platform for used-tractor dealerships: inventory and lead routing, a price predictor, condition certification and an expense tracker, with an auction portal alongside. My part in this one was mostly the infrastructure.

A telematics deployment for Mahindra’s Swaraj Division, with its own tracking and computation services. The platform itself was already proven in AMK, so the work here was the infrastructure to run it at scale and the interfacing with new IoT hardware.

A telematics deployment for Mahindra’s Farm Division — the analytics on top of it, and a growing share of the cloud infrastructure beneath.